Picture a Ch'in army camped on a plain. It is not one camp. It is a grid of camps. The Central force, the Left, the Right, the Forward, the Rear — each one walled off from the others, each one ringed by its own temporary rampart, with no road running between them that a stranger could walk.1 Inside that grid the general has his own fenced patch, and so does the regimental commander, and so does the company commander down at the level of a hundred men.2 A soldier who belongs to one company of a hundred cannot stroll into the next. If he tries, the officer there is supposed to kill him on the spot — and if that officer lets him live, the officer shares his crime.2
That is the whole idea before any of the vocabulary arrives. An army that wants to keep its secrets does not start with codes or invisible ink. It starts by making it physically impossible for a person to be in a place he has no business being. The wall keeps the enemy out. The inner walls keep your own men from knowing more than their share. The defensive face of intelligence is plumbing and fences first, cleverness second.
This is the ancient Chinese practice of perimeter security and internal movement control — what a modern handler would call operational security and compartmentalization rolled into one. The military classics treat it as one of three concrete measures that follow from the general demand for secrecy: strict constraints on movement, secret communications, and deception.3 This page is the movement half. It is the apparatus that turns "be secretive" into a checkpoint a man either passes or dies at.
Sun-tzu sets the frame at the moment of mobilization: "On the day the government mobilizes the army, close the passes, destroy all tallies, and do not allow their emissaries to pass through."4 Wei Liao-tzu sharpens it to a counterintelligence aim — "When itinerant persuaders and spies have no means to gain entrance, this is the technique for rectifying discussions."5 The point is not abstract tidiness. Sawyer notes that the historical record shows spies kept getting into encampments and walking out with vital information anyway, which is precisely why the regulations had to be this brutal.6
The late Ming Ping-fa Pai-yen gathers the attitude under one word: "caution." Its definition is almost comically plain — "Being careful in every aspect is termed 'caution.'"7 But the elaboration is where the texture lives: "Enter the army to assume command as if there were spies all about; when crossing the border, be as strict as when approaching battle. Whenever you obtain or seize something, make certain no harm is entailed; when you encounter natural obstacles of terrain, you must search about for villains."7 Terrain that hides things must be assumed to hide a person. A captured object must be assumed to carry a trap.
The mechanism is a chain of boundaries, each with a gate, each gate requiring a credential. Walk the chain.
Along the roads that crisscross the camp, administrative posts stand every 120 paces, sited so that each post is within sight of the next.8 Crossing the roads is forbidden; the roads are kept clear. A soldier moving through must carry a tally or token issued by a general or commanding officer — no tally, no passage.8 Foragers, fodder-seekers, herders move only in squads of five; a man alone, or a group that is not a five, cannot cross.8 At the crossing gates the guards execute the officer who has no token and the soldiers who are not in their five. Anyone who oversteps the demarcation lines is killed.8 The logic the text states explicitly: if inside the army no one breaks the prohibitions, then outside there will be no infiltrator who is not caught.8 The internal discipline is what makes the external screen reliable. A stranger stands out only when every legitimate man is accounted for.
Two outer layers extend the grid past the wall. The Sung Hu-ling Ching puts observation towers about a hundred paces outside the camp to watch for enemy movement and signal its presence and direction by flag — reporting force size, type, activity, and speed.9 Then come roving patrols, drawn from "courageous soldiers knowledgeable about the mountains, rivers, springs, and wells," each patrol confined to a designated zone well outside the camp, tasked with "seizing living spies dispatched by the enemy."10
The detail that makes this page worth reading is what the texts do to their own patrolmen. The roving patrols are kept deliberately ignorant of their own army's troop dispositions and defensive measures.11 The reason is cold and exact: if an enemy reconnaissance party captures one of them and interrogates him, he has nothing to give up. He cannot betray a layout he was never shown. The need-to-know principle is not a courtesy here. It is a survival calculation run on your own men — you blind the asset you are most likely to lose.
Compartmentalization as a doctrine, not an instinct. Most of the vault's intelligence pages are about acquiring foreknowledge — running spies, reading defectors, corroborating reports. This page is the inverse posture: assume the enemy is doing all of that to you, and architect your own army so that no single captured node yields the whole picture. It hands Systematic Covert Programs its defensive twin — a covert program needs not only collectors but a perimeter that prevents the other side's collectors from reading it.
It also supplies the physical substrate that the communications pages assume. Secure Military Communications describes watchwords and signal fires; those only matter inside a perimeter where unauthorized people genuinely cannot move. The fence and the cipher are one system. And it gives Field Observation Methods its mirror: the same scout-craft that reads the enemy's camp is what the towers and roving patrols are built to defeat when the enemy reads yours.
Take the single sharpest move in the whole passage and turn it over. The Hu-ling Ching wants roving patrols outside the camp to catch living enemy spies.10 To do that job well, a patrolman should be experienced, local, observant — he should know the springs and the ridgelines.10 That same competence makes him a high-value capture for the enemy. A skilled patrolman knows where the camp is soft. So the doctrine does something that looks self-defeating until you sit with it: it withholds from the patrol the very information that would make them effective if they were honest collectors for their own side. They are not told their army's dispositions or defenses.11
The trade is explicit. The patrol loses the ability to integrate what it sees with the camp's posture — it cannot think "the enemy is massing where we are weakest" because it does not know where the camp is weak. In exchange, the camp loses nothing when the patrol is taken. The information value the patrolman would have carried is set to zero in advance. This is the logic that later intelligence services call source compartmentation: you do not give a man more than his task requires, precisely because his task may end with him in an interrogation chair. The ancient text reaches this conclusion without any of the modern vocabulary, by the plain arithmetic of "what can this captured man tell them?"
It is worth holding the discomfort here rather than smoothing it. The doctrine treats its own loyal soldiers as future leaks to be pre-emptively sealed. The same word — caution — that tells a general to search terrain for villains also tells him to handle his own patrolmen as if they were one capture away from betraying him.
You are a gate guard at a crossing post on the camp road, the sun going down. You can see the next post's torch 120 paces off; if it goes dark, you will know something is wrong before any shout reaches you.8 A man approaches. He is in your army's uniform. That means nothing — the regulations exist because men in the right uniform have walked in before.6 You do not look at his face. You look at his hand. Tally or token, issued by a named officer, or he does not pass.8 He has none. You also notice the three men behind him are three, not five. A foraging detail moves in fives or it does not move.8
There is no negotiation in the procedure. The text does not give you a discretion to exercise; it gives you a knife to use and a punishment if you do not.8 You are not deciding whether this man is a spy. You are deciding whether he satisfies the credential, and the credential is the only thing you are permitted to read. Somewhere a general has already decided, by building this grid, that the cost of killing the occasional confused friendly is lower than the cost of one enemy walking the camp at night. You are the edge of that decision. You quiet your voice and your breath, as the Ping-fa Pai-yen says the cautious ancients did, and you watch the demarcation line, because a man who oversteps it has already answered the only question that matters.12
It fails the way every perimeter fails: at the seam between the rule and the man enforcing it. The Ping-fa Pai-yen names the failure before it happens — caution arises from doubt, from numerous doubts, and where doubt slackens, so does the search.13 A camp that grows comfortable stops searching the terrain for villains. The towers are manned by men watching the horizon out of habit, no longer reporting size, type, and speed because nothing has happened in a week.
The deeper failure is internal. The regulations work only if internal discipline is total — "if within the army no one contravenes orders, then without there will not be any that are not caught."8 Loosen the squads-of-five rule for foragers because it is inconvenient, let officers cross the roads because they are officers, and the stranger no longer stands out. The screen depends on every legitimate movement being predictable; the first exception you grant your own men is the gap the enemy uses. Sawyer's own caveat is the loudest warning sign: despite all this apparatus, spies still routinely got in.6 The presence of the machinery is not proof it is working. A camp can run every checkpoint and still leak, because the man at the gate has stopped looking at hands and started looking at uniforms.
The texts span centuries — the Six Secret Teachings is early, the Wei Liao-tzu preserves Ch'in regulations, the Hu-ling Ching is Sung, the Ping-fa Pai-yen late Ming.14 Sawyer presents them as a continuous tradition of camp-security thinking, but that continuity is his synthesis, and the gap between a Warring States precept and a Ming definition is large enough that independent verification is warranted before treating them as one doctrine. They agree in spirit; they may not descend from each other.
The honest tension is between the regulations and reality. The whole reason these draconian rules exist is that the softer version did not hold — spies kept getting through.6 So the texts document an aspiration as much as a practice. Did camps actually execute foragers who broke the five? The regulation says so; the historical writings that show spies succeeding suggest the rule was honored unevenly. The page should be read as the doctrine's maximal demand, not a description of a sealed army.
One open question the source leaves: the blinded-patrol logic assumes a captured man will talk under interrogation, which is why he is given nothing to tell. But it never asks what the enemy learns from the fact of capturing a patrol that knows nothing — an interrogation that yields nothing is itself a signal about how the camp is run.
Sawyer reads the movement-control texts as a coherent OPSEC tradition, and on the level of attitude they do converge — every source treats the camp as porous until proven otherwise. But the sources pull against each other on where the danger lies. The Wei Liao-tzu's Ch'in regulations locate the threat at the perimeter and answer it with execution at the gate: keep the wrong body out.8 The Hu-ling Ching locates it in the captured insider and answers it with enforced ignorance: assume your own man is taken and give him nothing.11 Those are not the same fear. One distrusts the stranger; the other distrusts what loyalty cannot protect against — torture. Sawyer files them together, but the patrolman who is blinded by his own side lives in a darker imagination of war than the gate guard who simply checks a tally.
The Ping-fa Pai-yen converges them under "caution" and in doing so almost dissolves the distinction — its definition stretches caution from terrain inspection to the handling of one's own associates.7 That breadth is its strength as a unifying frame and its weakness as a guide: when one word governs both the wall and the patrolman's ignorance, it stops telling you which threat you are actually defending against.
Handshake with Secure Military Communications. These two are the front and back of the same shield, and reading them together exposes a dependency neither states alone. Secure communications give the army watchwords, signal fires, and trained couriers — but every one of those assumes the courier can be physically isolated from enemy hands and that the watchword cannot be overheard by someone who shouldn't be present. Movement control is what makes that assumption true. The cipher is only as strong as the perimeter that keeps the enemy from standing close enough to read the signaler's flags. Conversely, the perimeter is only as useful as the communications that let a sealed-off segment still respond to the rest of the army. The insight the pair generates: an army can have perfect secrecy and lose anyway, because total compartmentalization severs the responsiveness that lets segmented forces recombine — the wall that protects you can also strand you. The Ping-fa Pai-yen names exactly this danger on the communications side, that segmented forces unable to recombine follow "the Tao of defeat."15 The two pages, held together, show that OPSEC is a tuning problem, not a maximization problem: more wall is not always more safe.
Handshake with Systematic Covert Programs. A covert program is an organization built to penetrate the enemy. Camp security is the same organization viewed from the target's side. The structural parallel is exact: both rely on need-to-know, both compartmentalize so that one captured node does not unravel the whole, both treat their own people as potential leaks. The tension that produces the insight: a state that runs a sophisticated covert program knows exactly how its own spies get inside enemy camps — and that knowledge is precisely what should make it paranoid about its own perimeter. The blinded patrolman exists because the people who designed him also ran the spies he is built to catch. Counterintelligence is not a separate discipline from intelligence; it is intelligence turned to face inward, designed by people who know the offense too well to trust their own fences. Neither page generates that on its own — the offensive page assumes a porous enemy, the defensive page assumes a porous self, and only together do they show that the same expertise produces both the lockpick and the lock.
Sharpest implication. The most secure node in the system is the one that knows nothing — the patrol blinded on purpose. The ancient answer to "how do I keep my men from betraying me under torture" is not better loyalty or better torture-resistance; it is to ensure there is nothing to betray. Security migrates toward ignorance. The most trustworthy courier is the one who cannot read what he carries.
Generative questions.